A/B testing is an evaluation method that consists of testing in parallel two different versions of the same product, to decide which one is more suitable for a specific user need or business goal. It can also be used to assess a new version of an existing product against its predecessor. It is best suited to testing small, incremental changes in a design solution. By restricting changes to a known variable, it is then possible to understand the effect of that variable.
Continued on pages 20–21 of Design. Think. Make. Break. Repeat.